Well-founded fears of some fowl times yet to come

Bird flu might not turn into the much-feared pandemic immediately, says Neil Ferguson, but we should not be complacent

To describe the cover of this book as sensational is an understatement. Above the tabloid title, a malevolent chicken fixes a gleaming eye on the reader. But despite the elements of a B-movie-style "invasion of the killer chickens" tale to this book, much of the narrative is engaging and often provocative.

The Monster at Our Door tells the story of the emergence of the H5N1 "bird flu" strain that has now become endemic in Asian bird populations and discusses the risk that H5N1 will start another human flu pandemic. By way of a brief review of influenza virology and the 1918-19 Spanish flu pandemic, Mike Davis makes a passionate case that the current crisis is at least in part man-made, and that our response to it thus far is wholly inadequate.

The author is a man with a mission, and readers seeking a dispassionate layperson's account of the H5N1 bird flu crisis will need to look elsewhere. This book is highly polemical, particularly in arguing that globalised agribusiness plays a role in accelerating the emergence of new influenza strains.

As the weekly recipient of a locally sourced organic vegetable box, I share many of the social and animal welfare concerns about industrialised large-scale livestock production. Predatory international corporations of the type that dominate global agribusiness make easy targets for those with an anti-globalisation bent. But it is far from clear that the move towards industrial production really increases the risk of a human influenza pandemic - or even that it has played an important role in enabling H5N1 to become entrenched in Asia.

Without doubt, there has been gross under-reporting of bird flu outbreaks across Southeast Asia, at least in part due to commercial self-interest, but then such self-interest operates at all scales of farming. A key problem in China until recently has been the partial compensation paid to small farmers whose poultry is culled as part of disease control efforts.

Who would want to be the village farmer responsible for all his neighbours losing their livelihoods? Even where full compensation is paid, farmers still underreport disease occurences: my own work suggests that UK farmers in the 1990s may have reported only one in five BSE cases in cattle.

From a purely epidemiological viewpoint, it is arguable that industrialised production actually reduces the risks of disease spread and cross-species transmission. First, because the commercial impact of a disease outbreak can be huge in large, high-density units, they tend to have better biosecurity than smaller units. Second, if disease does take hold, the resulting epidemic can be swift and devastating, but the opportunities for people to be exposed are arguably much more limited than they are in traditional "back-yard" production, where almost every family in a rural community keeps poultry and would be exposed to avian flu viruses in the event of an outbreak.

The author cites the example of the H7N7 avian influenza outbreak in Holland in 2003 as evidence for the role that industrialised poultry production might have in worsening an outbreak. However, although the economic cost of that outbreak was accentuated by the impact on large production units, the spread of the disease occurred principally between neighbouring small-scale free-range flocks.

In a similar vein, The Monster at Our Door claims that increasing human population sizes and urbanisation - in particular, high-density slums - offer "perfect environments for the evolution of flu virulence". Quite apart from the fact that virulence (the lethality of an organism) and transmissibility are here being confused, there is remarkably little data to suggest that the rate of spread of respiratory diseases such as flu varies much with population density. Much more important is the number of close habitual contacts - family members, classmates, work colleagues - and these vary much less than population density. Family size is arguably a risk factor for emergence, as initial clusters of cases are likely to occur within the close-contact environment of a household. Paradoxically, household sizes are larger in traditional rural communities than they are in the rapidly increasing megacities of China and Southeast Asia.

That said, I enjoyed many aspects of the book. Davis's review of the 1918-19 pandemic and why mortality was apparently much higher in India than in Europe is excellent. And I entirely agree that today in Africa "Aids might become influenza's deadly dancing partner, like malnutrition in India or malaria in Iran in 1918". The discussion of the likely mortality from an H5N1 pandemic is also refreshingly honest; Davis is right that the baseline figures quoted by the World Health Organisation and many governments (including our own) are certainly overoptimistic. Frankly, other than historical precedent, which is notoriously unreliable, we have no data to support the frequently made assumption that for H5N1 to become transmissible from person to person, it would have to become much less lethal. In reality we just do not know how lethal an H5N1 pandemic might be.

Stylistically, the book juxtaposes a dense narrative (for example, the first chapter's discussion of influenza biology and genetics) with occasional bouts of tabloid-ese. This can be, by turns, irritating and engaging. The prose reads a little like an extended Sunday supplement piece - and it appears to have been written in haste, to judge from the sometimes erratic referencing.

Davis concludes with an indictment of US infectious disease research funding policy - claiming, largely justifiably, that large sums were misdirected towards bioterrorism defence research and away from the more real threat of influenza and other emerging pathogens. The author claims that the response of the US Government, in particular to the threat posed by H5N1, has been wholly inadequate. In saying this, the book is now slightly dated: irrespective of the merits of the US response before August 2005, Hurricane Katrina has helped to catapult pandemic preparedness to the top of the US Government's agenda. Some $7 billion (£4 billion) has now been committed to developing vaccines, buying antivirals and enhancing global surveillance.

However, the wider indictment of the decline in public-health budgets in the US is more justified, and indeed could be extended to many other developed countries. Whether we have the infrastructure to mount an effective response to a pandemic is far from clear. The type of responses being envisaged - the delivery of drugs to millions of people per week during an epidemic - has never been attempted. And of course, as Davis justly points out, the medical responses that may be affordable in the West will protect at most 20 per cent of the world's population. The usual gap in equality of medical provision will be even more apparent in a future flu pandemic, where access to scarce and rationed vaccine and drug suppliers is controlled by wealthy producer countries.

Many people - perhaps including Davis - will argue that the responses that the US Government and others have announced in the past few months may be if not too little then too late. However, this presupposes that we face an imminent threat. In this, I would disagree with both the conclusions of this book and indeed with the current consensus of scientific opinion.

I question whether there are any robust data to support the assumption that an H5N1 bird flu epidemic increases the risk of a human pandemic. All we know is that birds are infected with many strains of flu all the time, and that every few decades a strain jumps from birds to people to cause a new pandemic. The conditions needed for that jump are not well understood. The continuing human and animal cases of infection with H5N1 are cited as evidence of pandemic potential. However, we have no idea how many people are infected each year with so-called low-pathogenicity avian flu strains, which have almost no symptoms. In the 2003 Dutch outbreak of highly pathogenic H7N7 avian flu, arguably more people were infected than during the whole of the global H5N1 outbreak. Most of these infections were only mildly symptomatic or, indeed, entirely asymptomatic - and so would not have been picked up had it not been for the intensive human testing programme undertaken in that outbreak. Because it is the number of human infections with a particular human flu strain that determines the risk of the virus evolving to become transmissible, it is far from clear whether H5N1 - despite its high public profile - is the strain that poses the highest risk of starting the next pandemic.

However, I wholeheartedly agree with one of the key messages of this book, namely that if H5N1 were to cause the next human flu pandemic, the consequences could be catastrophic, particularly for the bulk of the world's population with no access to drugs or vaccines. That alone justifies this book's call to arms.

In my view, the odds are that there will be no human pandemic in the next five years. However, this is no reason for complacency or a reduction in preparedness planning. As the emergence of a pandemic is, in essence, a chance event, we cannot predict exactly when the next pandemic will strike or how severe the consequences will be. That it will strike eventually is certain. I would therefore argue that we need to be prepared not just for the immediate (perhaps unlikely) threat, but for the long term. If nothing happens in the next few years, it will not have been because the risk was not there - but because we have been lucky.

Neil Ferguson is professor of mathematical biology, Imperial College London, and specialises in infectious disease epidemiology. He advises the UK and US governments on influenza pandemic preparedness.